Accurate and Memory-Efficient GPU Ray-Casting Algorithm for Volume Rendering Unstructured Grid Data

Abstract

We present a novel GPU ray-casting algorithm for volume rendering unstructured grid data. We employ the per-pixel intersection list to guarantee accurate results for non-convex meshes. For efficient memory access for the lists on the GPU, we represent the intersection lists for all faces as an array. To increase ray-coherency in a thread block and improve memory access efficiency, we propose an image-tile based ray distribution method. Also, we found that a prior approach using a per-thread local buffer to reduce redundant computation is not proper for recent GPUs. Instead, we use an on-demand calculation strategy that achieves much better performance even when it allows duplicate computation. With a GPU, our method achieved up to 36.5 times higher performance for the ray-casting process, and 18.1 times higher performance for the entire volume rendering, compared with the Bunyk algorithm using a CPU core. Also, our approach showed up to 8.2 times higher performance than a GPU-based cell projection method while generating more accurate rendering results.